SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
Using WITNESS™ Simulation Software as a validation tool for an industrial plant layout ENRICO BRIANO CLAUDIA CABALLINI ROBERTO MOSCA ROBERTO REVETRIA
DIPTEM – Department of Industrial Production, Technology, Engineering and Modelling Via Opera Pia 15, Genoa, ITALY
[email protected];
[email protected];
[email protected];
[email protected]; Abstract - Modeling and Simulation techniques are often powerful tools devoted to analyze the best layout for an industrial plant. In fact, these methodologies allow investigating the most suitable features or parameters regarding for instance buffer capacities or the number and type of machines and facilities, like for in-stance handling means. In literature, there are many examples of simulation models aiming at reproducing different plant layouts through the use of “What if” analysis. The focus of this work is on a plant layout simulation model developed by the Department of Production Engineering, Thermoenergetics and Mathematical Models of the University of Genoa, with the use of the commercial software WITNESS™, that enables the discrete event simulation. After the description of the model, the results obtained are presented and discussed, highlighting the importance of using this kind of simulation for layout design purposes.
Key Words – Plant Layout, Simulation, What if analysis, Logistics, WITNESS™ Software
propose an application of simulation for the design of
1. Introduction
an industrial plant layout with the goal of analyzing
The implementation of a new plant layout is a typical
flows, criticalities and bottlenecks, using WITNESS™
matter studied by logistics experts using Modeling and
software. This tool allows the formulation of “What If”
Simulation (M&S) techniques, and it is also one of the
analysis and animated simulations in two and three
major aspects discussed in academies to educate
dimensions.
young engineers and researchers about industrial
The paper is divided in the following sections: after this
engineering. There are in fact several aspects to be taken
into
account
during
the
design
introduction, which explains the aim of this research, a
and
short literature review is proposed in Section 2, while
implementation of a new plant layout, like distances,
Section 3 provides a short description of the model
buffer capacities (if present) and number and types of
developed in WITNESS™. Finally, Section 4 provides
machineries and ancillary facilities, like forklifts, human
some results and conclusion on the model developed.
re-sources and so on. For these reasons the authors
ISSN: 1792-507X
201
ISBN: 978-960-474-230-1
SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
authors have used another commercial software:
2. Literature Review
WITNESS™ from Lanner Group. WITNESS is a
The plant layout design is a typical matter that can be
process simulation software commonly used for both
approached with a modeling and simulation approach.
educational and business purposes, and it is one of the
As a matter of fact, in literature, there are many
most suitable for simulating plant layouts. In different
examples of simulation models devoted to analyze this
real business cases this tool has been successfully
topic, and most of them have been implemented using common
commercial
software
like
Simul8™
used to design and validate plant layouts; in the recent
or
past this software has been used to support the
Arena™. Smutkupt and Wimonkasame (2009) utilized
European aeronautics company Airbus in de-signing
Arena to develop a model that overtakes the limited
the wing production facility for the new passenger
results provided by the implementation of the CRAFT
aircraft A380, the world’s largest airplane ever
(Computerize Relative Allocation Facilities Technique)
produced. Airbus used WITNESS to ensure a lean
methodology, by showing more information of the
wing production process and to provide a good overall
design layout like the total time in system, waiting time
view of the factory life. Moreover the model was able to
and utilization. This is made possible thanks to the
give a detailed perspective about all the process
simulation model developed in ARENA™ which it is
phases. Finally, he layout designed using WITNESS
properly connected to the design system and to an
brought significant savings in terms of equipment.
output report system using ad hoc Microsoft Visual Basic™ interfaces.
3. The WITNESS™ Simulation Model
Simul8™ is used in Magoulas et al. (2002) for the design of a steelworks facility. This stochastic model
This section provides a description of the model
allows handling the uncertainties that arise from
implemented by the authors for the proper design of an
stochastic elements in the environment and in the
industrial plant layout. In particular the simulation
objective function evaluation process.
model developed reproduces a widget production line,
The authors themselves developed, in (Briano et al.
where different flows of parts and materials are
2004) an easy SIMUL8 model devoted to analyze the
managed using just one normal forklift of the capacity
performance level of a dental technician lab.
of 2000 kilograms. The objectives of the simulation
Taking into account the wide literature regarding the
model are to identify any criticalities in the system in
layout design, the authors propose the use of
order to validate an hypothetic plant layout considering
simulation in order to validate a widget plant layout
a certain daily flow of materials, to dimension inter-
scenario. More specifically, the simulation is used to
operative and machine buffers and to evaluate the
evaluate the utilization coefficient of the forklift truck
saturation
used for handling the various parts, for determining the
productive machines. The model allows formulating
average and maximum number of parts in queue
“What if” analysis in order to identify eventual
waiting to be processed and the average WIP (Work in
bottlenecks on the system. Finally the model has been
Progress) value and time. In order to do this, the
used in academy to train young engineers and
ISSN: 1792-507X
202
coefficient
of
the
different
types
ISBN: 978-960-474-230-1
of
SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
researchers to become confident with stochastic
•
simulation and “What if” analysis.
up a part;
In detail, the model reproduces a widget plant in which
•
different flows of materials are interfaced. Each one of
searching for a mission;
this flow feeds a series of different machines devoted
•
to transform the semi finished products in a finished
machine or a buffer.
widget ready to be sent to the shipping warehouse;
All the status are reported both in a tabular and a
each flow has a buffer, near to the referring workplace,
graphical form, showing a chart states for the objects.
where the queuing different parts wait before being
Buffers have also characteristic parameters, in terms of
processed.
capacity, which can be set by the user, in order to fix
In WITNESS software it is possible to set characteristic
an upper limit to the parts entering the system. It is
parameters for machines, forklifts and buffers, like the
possible to see significant output parameters like the
capacity, the processing times or the speed (for the
WIP, both in terms of time and number of parts, or the
forklift), allowing “What if” analysis on different
average and the maximum number of parts in queue,
scenarios. The model developed presents a unique
in order to identify criticalities and bottlenecks of the
forklift serving all the machines and the flows inside the
system.
plant, including also the possibility to take semi finished
However, several of the input parameters have to be
products from outside, moving along predefined
set in the track object: here it is possible to insert the
segments which connect machines, buffers and entry
loading/unloading time of the forklift, the real length of
points; the logic behind the forklift picking sequence is
the track and its capacity (in case of more than one
a “taxi” one, where the forklift presents a list of
forklift); in the same object it is inserted also the
missions to be satisfied and added in the queue once a
sequence logic of the forklift mission with the SEARCH
buffer or a machine “calls” the forklift. In the unloading
function.
procedure it is foreseen to set “parking points” for the
WITNESS allows implementing different logics for the
forklift devoted to hold it when it is idle along some of
parts assignment to the machines, making possible the
the segments.
use of specific functions and statements, like FOR
Forklifts and machines in WITNESS can assume
cycles, IF statements and so on. In particular, in the
different status:
model developed by the authors, the WRAP function
Demand: the forklift is in mission going to pick Transfer: the forklift is moving along the tracks Loaded: the forklift is carrying a part towards a
has been used, which assigns randomly the parts to •
Idle: the object is inactive;
the first free machine and then follows a sequential
•
Busy: the object is working;
feeding configuration. WITNESS allows also input
•
Blocked: the object is not able to manage the
initializing, in order to start the simulation with a
missions due to the high workload.
particular configuration of busy machines, editing a text file (extension “.sta” ) which contains the desired
But there are other status typical of the forklift:
ISSN: 1792-507X
parameters.
203
ISBN: 978-960-474-230-1
SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
component for the final widget. Setting the parts’ interarrival time for each machine of each series, the saturation coefficients for the forklift and the machines themselves have been analyzed, in order to evaluate
4. Conclusion
the presence of bottlenecks and the feasibility of the plant layout using just one forklift. The WITNESS™
The WITNESS model has been developed in order to
software allows analyzing the results of the simulation
evaluate the feasibility of serving all the different machine
types
with
only
one
normal
in two different ways: both in a tabular form,
forklift,
considering the percentages of the simulation time in
considering a certain inter-arrival time between the two
the different status seen in the previous section, and in
parts for each machine. Six different types of machines
a graphic one, underlining the different status in
transforming each one a different raw material part in
different colors.
semi-finished products have been considered, as
Figure 2 represents the status chart for the forklift in
represented in Figure 1, that represents the conceptual
the base case scenario, considering an inter-arrival
model of the production flow with the six different parts
time for Part1 of 300 minutes.
feeding the productive machines working in parallel. The semi finished products, stored in their relative buffers, are then assembled in the final assembling machine and then sent to the shipment warehouse.
Figure 1: The conceptual model In particular the attention has been focused on the most critical flow referred to Part1, which is the main
ISSN: 1792-507X
204
ISBN: 978-960-474-230-1
SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
Table 1 shows how the forklift is saturated for more than the 97% of the time, but it is always in transfer
Figure 2 – Chart Status for Forklift Saturation
from one point to another of the plant because it is not The graph represented in Figure 2 shows that the
able to satisfy all the missions. For the same reason
forklift for the most of the time is in transfer in the plant
the parts, especially Part1, accumulates in queue in
(that means that is moving to different points of the
the buffers; this implies that the machines are idle for the most of the time, as shown in Figure 3, representing the Chart States of Machines M11-M14 trans-forming Part1 in semi-finished products.
plant called by the various operators), it is idle (parked) for a little time, and it carries parts for a short time. This implies that in the buffers significant queues are generated. This means that the warehouse is definitely not optimized. The percentages in detail are presented in Table 1. Figure 3: Chart Status for M11-M14
Table 1: Forklift Saturation Percentages
The numerical results are reported in Table 2:
Name
Forklift
% Idle
2.86
% Demand
1.71
% Transfer
93.93
Name
M14
% Loaded
1.49
% Idle
95.83 97.22 97.92 97.22
% Stop
0.00
% Busy
4.17
2.78
2.08
2.78
% Blocked
0.00
% Filling
0.00
0.00
0.00
0.00
Distance
142747.19
% Emptying
0.00
0.00
0.00
0.00
Loads
613
% Blocked
0.00
0.00
0.00
0.00
ISSN: 1792-507X
Table 2: Status Percentages for M11-M14
205
M13
M12
M11
ISBN: 978-960-474-230-1
SELECTED TOPICS in SYSTEM SCIENCE and SIMULATION in ENGINEERING
% Cycle Wait Labor
0.00
% Setup
0.00
0.00
0.00
0.00
% Setup Wait Labor
0.00
0.00
0.00
0.00
% Broken Down
0.00
0.00
0.00
0.00
% Repair Wait Labor
0.00
0.00
0.00
0.00
No. Of Operations
6
4
3
4
0.00
0.00
0.00
References [1] Smutkupt U., Wimonkasame S. 2009. Plant Layout Design with Simulation. In Proceedings of the International MultiConference of Engineers and Computer Scientists 2009 Vol II IMECS 2009, March 18 - 20, Hong Kong [2] Magoulas G.D., Eldabi T., Paul R.J. 2002. Adaptive Stochastic Search Methods for Parameter Adaptation of Simulation Models. In Proceedings of 2002 First International IEEE Symposium “Intelligent Systems”, pp 23-27, September 2002. [3] VV.AA. 2010. Health-Care Simulation. Available via http://www.visual8.com/healthcare_simulation.html [accessed March 10, 2010] [4] Briano E., Mosca R., Revetria R., Savastano F. 2003. Integrating Wireless Technologies and Data Warehousing for Distributed Imaging Based Support in an Orthodontic Practice. In Proceedings of IASTED BIOMED 2004, Innsbruck, February 16-18 [5] Lanner Case Study. 2003 WITNESS helps Airbus perfect plans for massive A380 wing production facility. Available via http://www.dynamic.co.kr/Witness_Training_Center/Ca se/Airbus%20A380.pdf [accessed March 10, 2010].
Table 2 shows that the machines M11-M14 are not working for the most of the time, because the parts are queuing in the buffers; this is confirmed also by the number of the operation processed by each machine. In conclusion, the forklift is sufficient for the parts handling inside the plant, but there is no doubt that an optimization
of
the
handling
process
must
be
implemented: this is possible by increasing the interarrival time of the parts or reducing the distance between the warehouses and the plant itself. The segment length is in fact one of the most significant variable to improve the system performance reducing the
saturation
coefficient
of
the
forklift
and,
contemporarily, increasing the saturation coefficient of the different machines types. Finally, it can be stated that WITNESS provides a powerful tool to implement simulation models de-voted to analyze plant layouts and to perform feasibility analysis; in particular, it supports decision makers in formulating “What If” analysis before performing significant investments like a new plant layout or the purchase of expensive handling means.
ISSN: 1792-507X
206
ISBN: 978-960-474-230-1
